[Ffmpeg-devel-irc] ffmpeg-devel.log.20141124

[burek]

[01:01] <cone-717> ffmpeg.git 03Michael Niedermayer 07master:8294f5042d5c: Changelog: remove 24bit lossless wma entry for now, some samples work others do not yet work
[01:20] <BBB> michaelni: do you still do daily merges from libav?
[01:20] <BBB> I have doubt about one of their patches from today
[01:21] <michaelni> BBB, yes, which patch? i see no new commits in libav
[01:22] <BBB> the vp9 one
[01:22] <BBB> it might not have been pushed, I see it on the list
[01:23] <michaelni> whats the subject of the mail/patch ? and what should be done with it ? should i ignore it?
[01:23] <BBB> [libav-devel] [PATCH 08/10] vp9: Use the correct upper bound for seg_id
[01:24] <BBB> Id ignore it
[01:24] <michaelni> i merged the #define and ignored the rest a few days ago
[01:24] <BBB> oh ok, good, thanks
[01:24] <BBB> the define is fine yes
[01:25] <BBB> oh right I see its 2 days old
[01:25] <BBB> I dont check their list very often
[01:25] <BBB> thanks
[01:25] <michaelni> np
[01:28] <rcombs> heh, just glanced at their git and list to see if some DASH patches I'm waiting for are in yet (they're not) and was surprised to see one of my commits
[01:28] <rcombs> (properly attributed, mind)
[01:39] <michaelni> rcombs, if your work depends on / is blocked by some patches somewhere, you could include them in your patchset which you post to ffmpeg-dev. That assumes of course that these patches are "ready" and not unfinished
[01:44] <rcombs> michaelni: I think they'll be getting a little bit more work/testing on libav's end before getting merged there pretty shortly, and I don't need the changes urgently, so I figure I'll just wait and let the process proceed as usual
[01:44] <rcombs> these in particular, in case you're wondering: https://github.com/mstorsjo/libav/commits/dash-full
[01:45] <rcombs> I think mstorsjo's planning to replace that last commit with something more complete first
[02:05] <Compn> rcombs : wbs (mstorsjo) is here too
[02:30] <cone-717> ffmpeg.git 03Michael Niedermayer 07master:5182a2a235c3: avutil: remove FF_CONST_AVUTIL53, its no longer needed
[04:17] <cone-717> ffmpeg.git 03Lukasz Marek 07master:efe34e87ebf5: lavc/libxvid: fix mem leak in case of init failure
[04:17] <cone-717> ffmpeg.git 03Lukasz Marek 07master:02cb7d4c9c3a: lavc/smvjpegdec: fix mem leak in case of init failure
[04:17] <cone-717> ffmpeg.git 03Lukasz Marek 07master:c9d39fc8c687: lavc/huffyuvdec: fix mem leak in case of init failure
[05:14] <cone-717> ffmpeg.git 03Michael Niedermayer 07master:b4d8724ab28d: avutil/file: fix av_tempfile() documentation
[05:34] <rcombs> Compn: thanks, I'd forgotten his nick (derp)
[12:19] <cone-673> ffmpeg.git 03Vittorio Giovara 07master:9c12c6ff9539: motion_est: convert stride to ptrdiff_t
[12:19] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:ea41e6d63730: Merge commit '9c12c6ff9539e926df0b2a2299e915ae71872600'
[12:54] <cone-673> ffmpeg.git 03Vittorio Giovara 07master:065923b0781b: mpegenc: prevent a NULL pointer dereference
[12:54] <cone-673> ffmpeg.git 03Vittorio Giovara 07master:277ff7f5dc13: lavu: move internal define to the only places where it is used
[12:54] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:010adacbe236: Merge commit '065923b0781b06a2604f69f4e2c2407b7750a854'
[12:54] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:932d8d790ccf: Merge commit '277ff7f5dc134f1c2dfc4ea0ef3540340482e3d2'
[13:15] <cone-673> ffmpeg.git 03Vittorio Giovara 07master:b99ca863506f: aacdec: avoid an out-of-bounds write
[13:15] <cone-673> ffmpeg.git 03Vittorio Giovara 07master:30b8eb0f87b0: sol: simplify sol_codec_id()
[13:15] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:bcaef717ec0f: Merge commit 'b99ca863506f0630514921b740b78364de67a3ff'
[13:15] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:4fecf170d75d: Merge commit '30b8eb0f87b0eaefdc115ef38f8ad87dd3a6e50b'
[13:24] <cone-673> ffmpeg.git 03Luca Barbato 07master:299d8ab104fb: cook: Make sure there is enough extradata
[13:24] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:1db2d39dfdda: Merge commit '299d8ab104fb350254eb2e6d9ecdce892a2a55b1'
[13:37] <cone-673> ffmpeg.git 03Vittorio Giovara 07master:0562887a9843: tiffenc: initialize return value
[13:37] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:d2e0543766e0: Merge commit '0562887a984388fdc7a9b71c9374ff9c756fb4f1'
[13:46] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:55b59fab880a: roqaudio: Always use the frame buffer on flush
[13:46] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:f41d409b7a6c: Merge commit '55b59fab880a9fcdd30f97c5170af282087ac4f7'
[13:50] <wm4> ubitux: what happened to your patch that added a flag to demuxers whether a file is fully read?
[13:58] <cone-673> ffmpeg.git 03Vittorio Giovara 07master:208f3abb9177: aacsbr: always initialize max_qmf_subbands
[13:58] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:96398cc12332: Merge commit '208f3abb917757743313da0da714e525e03159d2'
[14:07] <cone-673> ffmpeg.git 03Martin Storsjö 07master:6f26f14f134e: sidxindex: Write mimeType=audio/mp4 for audio-only representations
[14:07] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:7a19a8fb696f: Merge commit '6f26f14f134e753d6168591f30815b1c08c1498b'
[14:15] <arpu> hi any openh264  patches around for ffmpeg? 
[14:15] <wm4> why would you want to use openh264 via ffmpeg??
[14:19] <kierank> patent fud probably
[14:19] <kierank> oh for encoding probably makes sense
[14:19] <kierank> it's openhevc that's not
[14:27] <arpu> wm4 for encoding only testing 
[14:40] <J_Darnley> A few people have asked that recently yet nobody seems to want it enough to provide a patch.
[14:41] <wm4> so, raw codecs don't output timestamps, but they're also not marked with AVFMT_NOTIMESTAMPS
[14:41] <wm4> is that as intended, or a bug?
[14:44] <cone-673> ffmpeg.git 03Martin Storsjö 07master:fe42f94ce102: dashenc: Don't segment all video streams when one stream gets a keyframe
[14:44] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:9a5730966522: Merge commit 'fe42f94ce1023f9c2f7e86404c60afcee5b078a9'
[17:29] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:1d242f9816ac: avformat/mpegenc: assert that premux_packet is non null
[17:29] <cone-673> ffmpeg.git 03Boris Reisig 07master:54170a33c2c9: avdevice/xcbgrab: fix undefined reference to xcb_shape_rectangles in xcbgrab.c
[17:54] <cone-673> ffmpeg.git 03Rodger Combs 07master:39f247121ec6: ffmpeg: fix accurate seeking with -copyts
[17:57] <wm4> michaelni: can I get access to coverity? is it enough if I "ask" for access on the coverity site?
[18:05] <michaelni> wm4, i just send you an invite
[18:06] <wm4> michaelni: thanks... seems I have access now
[18:06] <wm4> heh, there's not much
[19:42] <ubitux> wm4: so i was wondering if it was correct to close the fd
[19:42] <ubitux> because of later operations that could happen
[19:42] <ubitux> and i think that's why i didn't applied it yet
[19:42] <wm4> yeah, I was worrying about that too
[19:43] <wm4> it's orthogonal though
[19:43] <wm4> the avio layer probably needs a callback for this
[19:43] <ubitux> well you might not want to close it yourself either
[19:43] <ubitux> so i'm not sure if that patch in itself really helps
[19:45] <wm4> why wouldn't I want to close it?
[20:02] <ubitux> wasn't it the purpose of the patch?
[20:02] <ubitux> like being able to liberate fd etc?
[20:06] <wm4> ubitux: yes, but you said I "might not want to close"
[20:06] <wm4> what's the reason?
[20:37] <ubitux> wm4: the reason is that later in the framework we might want to access the pb again
[20:41] <ubitux> like you see the force flushing option
[20:41] <ubitux> for example
[20:41] <ubitux> but this might be just for muxing
[20:41] <ubitux> that's just an example
[20:41] <ubitux> i'm just not that comfortable doing that change yet
[20:43] <wm4> you don't need to include any actual closing
[20:44] <wm4> and the sub demuxers definitely don't read more data after opening
[20:45] <ubitux> can we be sure the framework isn't peaking for $randomreason?
[20:48] <wm4> well you just return EIO or EOF in this case, and probably nothing bad happens
[20:49] <wm4> but I coul check utils.c for possible issues
[20:49] <wm4> *could
[20:50] <ubitux> yeah see
[20:50] <ubitux> post read_header()
[20:50] <ubitux> you have an avio_tell()
[20:50] <ubitux> it's checking if the pb is set though
[20:50] <cone-673> ffmpeg.git 03wm4 07master:f41cf2e09ec2: avformat/rawdec: raw formats have no timestamps
[20:50] <ubitux> but well, might want to check a bit more
[21:03] <ubitux> damn, a v9
[21:04] <ubitux> i wonder when it's going to get upstreamed
[21:04] <ubitux> :D
[21:05] <rcombs> get hype
[21:06] <nevcairiel> i think he said he wanted to post a v9 with a important bugfix he had and then start splitting
[21:10] <ubitux> i think he said that for v7 or so
[21:16] <cone-673> ffmpeg.git 03Benoit Fouet 07master:9e1cfbd38a18: avformat/apngdec: transmit all the chunks between consecutive fcTL ones.
[22:10] <cone-673> ffmpeg.git 03Michael Niedermayer 07master:e6ea75c5e6b6: doc/examples/decoding_encoding: fix storing all channels
[23:06] <kierank> Gramner: any comments on the new v210 asm sets?
[00:00] --- Tue Nov 25 2014